1Naturally Wallace Consulting, P.O. Box 37, 126 2nd Street S, Stillwater, Minnesota 55082, USA
2Jacobs, 1295 Northland Drive, Suite 290, Mendota Heights, Minnesota 55120, USA
5.8.1 Introduction
Fill-and-drain wetlands are subsurface-flow wetlands that rely on the alternating filling and draining of wetland cells to move water (during the fill cycle) and air (during the drain cycle) in and out of the wetland cell. These systems are alternately called tidal flow or reciprocating wetlands (Austin, 2006;
Behrends, 1999). Fill-and-drain wetlands are a further development of the contact bed systems developed in late 1800s (Kinnicuttet al.,1910).
In the simplest configuration, a cell is filled using the influent flow and then drained so the next“batch”of water is then treated. Alternate configurations use recirculation so that the fill-and-drain cycle frequency can be adjusted independently of the influent flow rate.
Fill-and-drain wetlands are of interest because through the sequence of filling and draining, they cycle through aerobic and anoxic/anaerobic phases automatically. This makes them especially useful when removal of total nitrogen (TN) is a goal through nitrification/denitrification.
Fill-and-drain wetlands undergo cyclic changes in redox potential, ranging from aerobic (draining+ empty phase) to anoxic/anaerobic (filling+full phases). These cycling changes result in distinct treatment mechanisms operating in different phases of the treatment cycle:
• “Empty” Phase. Air, having been drawn into the bed during the draining phase, allows rapid oxygenation of biofilms (Behrends et al., 2001). Organic compounds having been previously adsorbed into biofilms are consumed by microorganisms under aerobic conditions, and positively charged ammonium ions (NH4+) are converted to negatively charged NO3− ions. Once the food supply is exhausted, further microbial activity results in the endogenous respiration, reducing the occurrence of clogging.
• “Filling”Phase. As the cell is filled, air is forced out of the system. Water enters first at the bottom of the treatment cell, and thus has the longest contact time. Chemical transformations of the“full”phase begin to occur.
• “Full” Phase. As the pore spaces fill with wastewater, oxygen is consumed and the redox potential decreases. Nitrate ions (NO3−), formed from previously oxidized NH4+, diffuse out of the biofilms into the bulk liquid. The presence of NH4+ (from the influent wastewater) and NO3−creates conditions suitable for alternate nitrogen transformations such as anammox. NO3−
can also serve as an oxygen supply for degradation of organic matter and for conventional denitrification; approximately 80% removal of total nitrogen (TN) has been observed (Austin et al., 2003).
Positively charged NH4+ions in the bulk liquid diffuse into the biofilms. This diffusion process and the overall total adsorption capacity of the bed is enhanced by the cation exchange capacity (CEC) of the bed media (Austin, 2006). Organic compounds are adsorbed into the biofilms, and this process is relatively rapid, taking approximately 5 minutes (Kinnicutt et al., 1910). When NO3− and dissolved oxygen are fully consumed, biodegradation of organic compounds can continue under anaerobic conditions.
• “Draining”Phase. Water is released from the bed. Rapid drainage times of 30 minutes or less are recommended, as this aids in drawing air into the bed (Dunbar, 1908). Chemical transformations of the“empty”phase begin to occur as the bed is drained, and the cycle begins anew.
5.8.2 Design considerations
Fill-and-drain wetlands are generally dimensioned based on clogging, hydraulics, and the number of fill-and-drain cycles per day, as summarized in Table 5.11. Flow is typically rotated through multiple beds in parallel or series, often using internal pumping to achieve the desired number of fill-and-drain cycles.
Once the wetland cells are dimensioned to avoid clogging, the most important design parameter becomes the number of fill-and-drain cycles per day, as this relates to oxygen transfer (Table 5.12). In many designs, this is related to a“rule of thumb”oxygen consumption rate of approximately 7–10 g O2/m3cycle (Wallace, 2014), with the number of cycles per day determined by the total oxygen demand (carbon+nitrogen) applied to the system. This“rule of thumb”is commonly used because it is simple, but it does not take into account the fact that oxygen is not limited during the “empty” phase, so ammonia removal is actually a function of the amount of ammonia adsorbed by the bed during the“full”phase.
Ammonia removal is related to the ammonia exchange capacity (AEC) of the bed materials, which determines the total amount of ammonia adsorbed during the “full” phase. The AEC is related to the presence of biofilms on the bed media and the cation exchange capacity (CEC) (ASTM D7503-18, 2018) of the material making up the bed media. Standard laboratory procedures for measuring AEC have yet to be developed, and designers typically devise tests specific to the project under consideration.
However, ammonia removal in fill-and-drain wetlands clearly improves when materials with a high CEC are utilized (Austin, 2006). Fill-and-drain wetlands with very high rates of ammonia removal have been designed and constructed based on AEC methods.
Overall performance (inlet vs. outlet) of fill-and-drain systems can be described using first-order kinetic rate coefficients (k) (Nivalaet al., 2019b), as summarized inTable 5.13. However, the diffusion processes Table 5.11 Typical design parameters for fill-and-drain wetlands.
Design Parameter Recommendation References
Pretreatment Primary treatment required Kinnicuttet al.(1910)
Influent loading (inlet cross-sectional area)
BOD5 ≤100 BOD5g/m2/d Wallace (2014)
TSS ≤100 TSS g/m2/d
Influent distribution ≤50 m2per feed point Dotroet al.(2017) Drainage system ≤30 min to drain bed (generally by siphon) Barwise (1899) Fill-and-Drain cycles 6–24 per day (6–12 per day common) Dotroet al.(2017)
Kinnicuttet al.(1910), Austinet al.(2003)
Media size 8–16 mm Kinnicuttet al.(1910),
Nivalaet al.(2014)
Number of beds 2–8 Nivalaet al.(2013a);
Austinet al.(2003) Treatment kinetics Table 5.13or pilot testing Nivalaet al.(2019b)
that occur during each phase of the treatment cycle appear to be rapid and occur more quickly than the beds can be physically filled and drained. As a result, each individual phase of the fill-and-drain treatment cycle have not yet been described using kinetics.
5.8.3 Potential design and operational issues
Loading fill-and-drain wetlands above recommended limits (Table 5.11) can result in clogging of the beds due to excess production of microbial biomass. Fill-and-drain wetlands are normally designed to receive primary-treated wastewater so that solids loadings on the beds are minimized. If primary treatment is problematic or is not provided, coarser bed materials (.75 mm) are required in the first treatment stage (Kinnicuttet al.,1910). The use of coarser bed materials to reduce the potential for clogging also lowers treatment performance, so more than one treatment stage is employed (Barwise, 1899; Dunbar, 1908), with the first stage essentially acting as a roughing filter.
Fill-and-drain wetlands operate with a variable water level, and the gravel used does not have the same capillary action as fine-grained soils. This can be an issue during vegetation establishment, especially in arid climates. When plants are fully established, the root systems will extend throughout the bed down to the minimum water level. When plant root systems are still shallow, they can lose contact with the water,
Table 5.12 Estimated oxygen transfer rates for passive and fill-and-drain wetlands (Wallace, 2014).
Wetland Type Estimated Oxygen Consumption (g O2/////m2/////d)
HF1 6.3
FWS1 1.47
VF (unsaturated)1 24.7 French VF (1ststage)2 40–60 Fill-and-Drain3 168–240
150th percentile values from Kadlec and Wallace (2009) assuming aerobic BOD removal and conventional nitrification.
2Data from France indicates that the first stage of French VF wetlands can sustainably operate at roughly 1.5 m2/PE (Molleet al., 2005).
3Roughly estimated at 7–10 g O2/m3per fill-and-drain cycle; at up to 24 cycles per day and a 1 m bed depth (Wallace, 2014). However, this also depends on the cation exchange capacity (CEC) of the media.
Table 5.13 P–k–C* model fit parameters for passive and fill-and-drain wetland systems at Langenreichenbach, Germany (Nivalaet al., 2019b).
Fill-and-Drain 2 0.3 672 2 0.1 450 2 4.4 123
increasing the risk of drought stress. This may require temporary irrigation systems during the plant establishment phase.
Treatment in the system is dependent on the number of fill-and-drain cycles per day. Designs that depend only on the influent flow rate to regulate the fill-and-drain cycle can have very slow cycling during low flows, consequently many designs use flow recirculation to increase the cycle frequency. Generally, it is best to design the system to support the maximum number of cycles desired, as it is much easier to slow down the cycle frequency than it is to speed it up.